Linear Layouts: Robust Code Generation of Efficient Tensor Computation Using 𝔽_2
Paper: https://
arxiv.org/pdf/2505.23819
.pdf
…
@jiqizhixin
-
Efficient Tensor Computation Code Generation Using Linear Layouts
By
–
-
OpenAI Linear Layouts: Tensor Mapping Innovation for DL Compilers
By
–
A paper from OpenAI? Long time no see. This paper introduces Linear Layouts, a general, algebraic formulation of tensor layouts using binary linear algebra over bit representations, addressing longstanding challenges in DL compilers like Triton. FYI, Tensor layout = mapping
-
Negative Reinforcement Boosts LLM Reasoning Effectiveness
By
–
The Surprising Effectiveness of Negative Reinforcement in LLM Reasoning
Paper: https://
arxiv.org/pdf/2506.01347
.pdf
…
Code: https://
github.com/TianHongZXY/RL
VR-Decomposed
… -
Penalizing Mistakes: Simple Yet Effective RL Reasoning Strategy
By
–
Just Say No: Penalizing Mistakes May Be All You Need in RL for Reasoning This paper, from UVA and Princeton, dives into Reinforcement Learning with Verifiable Rewards (RLVR) for reasoning tasks and reveals a surprising insight: penalizing wrong answers (NSR) can be more
-
Control-R: Controllable Test-Time Scaling for AI Models
By
–
Control-R: Towards controllable test-time scaling
Paper: https://
arxiv.org/pdf/2506.00189
.pdf
… -
Reasoning Control Fields Optimize Long Chain-of-Thought in LLMs
By
–
Controlling LLMs' Thinking Effort: Reasoning Control Fields (RCF) for Long CoT This paper tackles underthinking (too shallow) and overthinking (too verbose) in long chain-of-thought (L-CoT) reasoning by introducing Reasoning Control Fields (RCF) — a test-time method that
-
High-Entropy Minority Tokens Improve LLM Reasoning via Reinforcement Learning
By
–
Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
Paper: https://
arxiv.org/pdf/2506.01939
.pdf
…
Project： https://
shenzhi-wang.github.io/high-entropy-m
inority-tokens-rlvr/
… -
High-Entropy Tokens Drive LLM Reasoning in RLVR
By
–
It turns out you can drop 80% lowest-entropy tokens and still get good performance. This paper shows that only ~20% of tokens—those with the highest entropy—actually drive the reasoning process in LLMs trained with Reinforcement Learning with Verifiable Rewards (RLVR). The rest?
-
RLHF 101: Technical Tutorial on Reinforcement Learning from Human Feedback
By
–
RLHF 101 from ML@CMU A Technical Tutorial on Reinforcement Learning from Human Feedback "This blog dives into the full training pipeline of the RLHF framework. We will explore every stage — from data generation and reward model inference, to the final training of an
-
AlphaOne: Reasoning Models Thinking Slow and Fast
By
–
AlphaOne: Reasoning Models Thinking Slow and Fast at Test Time
Paper: https://
arxiv.org/pdf/2505.24863
.pdf
…
Project: https://
alphaone-project.github.io
Code: https://
github.com/ASTRAL-Group/A
lphaOne
… (coming soon)